1. 18
  1. 9

    This reminds me of the O(1) algorithm for solving the N-body problem with lightbulbs and photosensors. It was used in 1941 to do the first “simulation” of merging galaxies.


    The figure in the paper shows it produced nice tidal spiral arms:


    This works because both the strength of gravity and the brightness of a lightbulb fall of as 1/r^2.

    Every N-body simulation done on a computer is in some sense less efficient than this, since at best computer simulations use an O(Nlog(N)) algorithm.

    1. 8

      The best sorting algorithm is obviously sleep sort.

      1. 3

        There was a similarly wonderful sorting algorithm that was named something like Multiverse Sort, where on sort it spawned one process (one universe) for each permutation of the inputs and then every universe which was not “born” sorted is destroyed.

      2. 2

        Why is it that so many websites will switch you to a mobile link on mobile but none will switch you back on desktop?